Introduction:
Active surveillance (AS) is a well-established option in the management of low risk prostate cancer (PCa). The safety and efficacy of AS depend upon effective risk stratification at diagnosis and during follow up. Clinical characteristics at diagnosis have been shown to accurately predict initial risk of disease progression on AS. Studies have explored the correlation between AS outcomes and serial PSA measurement, biopsy findings, and imaging results during follow up. PSA density (PSAd) has been associated with risk of progression on AS when measured at diagnosis or confirmatory biopsy, yet no study has examined the correlation between AS outcomes and serial PSAd measurement during follow up. We investigated the prognostic impact of serial PSAd measurement (PSAd kinetics) in men enrolled in AS to evaluate whether PSAd kinetics independently predict pathologic grade progression on follow up AS biopsy.
Methods:
1290 men have enrolled in AS at our institution between 1997-2016 with their outcomes prospectively tracked. 453 patients with Gleason Grade Group 1 disease, ≥ two PSAd measurements at least one year apart during AS were identified. The primary outcome was pathologic grade progression on subsequent surveillance biopsy. A visual comparison of PSAd trends prior to this outcome or censoring was done using locally estimated scatterplot smoothing. The predictive power of PSAd kinetics was analyzed with a Cox proportional-hazards model and compared with a rich multivariable reference model of commonly used predictors at diagnosis (age, NCCN risk, gleason grade, disease volume and PSAd) using the concordance index (c-index) and likelihood ratio test. PSAd was used as continuous time-varying variable, reevaluating the risk of disease progression each time a new value was recorded. An extended Kaplan Meier estimator was used to graphically examine the risk of progression stratified by PSAd quartiles.
Results:
Median follow-up for the entire cohort of 453 men was 5.9 years (IQR 4.4-8.5). 137 men (30.2%) experienced pathologic progression on follow up AS biopsy at median time of 3.3 years after diagnosis (IQR 1.8-5.2) with 162 (35.8%) undergoing treatment. PSAd varied significantly during the course of AS with higher levels observed in patients who experienced pathologic progression (Figure 1). After adjusting for baseline covariates from the reference model, PSAd kinetics had the highest-ranking hazard ratio for grade progression [2.01, 95% CI: 1.72-2.36, p<0.001]. The risk of progression was strongest if patients had a PSAd value measured in the fourth quartile of PSAd values (>0.16 ng/ml/ml) with a distinct curve deviation and increased hazard ratio (4.25, 95% CI: 2.49-7.24, p<0.001) (Figure 2). Adding either PSAd kinetics or PSAd in quartiles to the reference model significantly improved the c-index of the cox regression (0.75 vs 0.69 respectively) indicating a better fit.
Conclusion:
Our data suggest that PSAd kinetics are a strong independent predictor of pathologic progression on Active Surveillance even after adjusting for baseline disease characteristics at time of diagnosis. In our cohort, PSAd values greater than 0.156 ng/ml/ml were associated with a markedly reduced progression free survival, thus should alert the treating physician to utilize close scrutiny and consider repeat biopsy. Since measurement of the components of PSAd (i.e. PSA and prostate volume) is already part of a routine AS protocol, monitoring PSAd over time confers no additional cost or burden to the patient and could be easily integrated into clinical practice to improve risk stratification for men with prostate cancer.
Funding: N/A
Image(s) (click to enlarge):
PSA DENSITY KINETICS PREDICT GRADE PROGRESSION ON PROSTATE CANCER ACTIVE SURVEILLANCE
Category
Prostate Cancer > Potentially Localized
Description
Poster #167
-
Presented By: Andrew Gusev
Authors:
Andrew Gusev
Florian Rumpf
Dimitar Zlatev
Jeffrey Twum-Ampofo
Douglas Dahl
Matthew Wszolek
Michael Blute
Keyan Salari
Adam Feldman